jgrusewski 74392e3744 fix(critical): 16-byte align weight pointers for cublasLtMatmul FAST_TF32 ...

Root cause: weight tensors packed sequentially in the flat params buffer
had non-aligned start offsets when preceding tensors had odd element counts
(e.g. bias of 51 atoms = 204 bytes, 204 % 16 = 12). cublasLtMatmul with
CUBLAS_COMPUTE_32F_FAST_TF32 requires 16-byte aligned buffer pointers.

Fix: pad each tensor to 4-element boundary (16 bytes) in both
f32_weight_ptrs_from_base (pointer computation) and compute_total_params
(buffer allocation). Added align4() and padded_byte_offset() helpers,
fixed shrink_perturb skip range and bottleneck gradient offset.

Switched compute type: CUBLAS_COMPUTE_32F → CUBLAS_COMPUTE_32F_FAST_TF32
(forward + backward). Explicit TF32 tensor core path, required by cuBLAS
13.0 on H100 SM90.

Deleted dead bf16_weight_ptrs function.
19/19 smoke tests pass on RTX 3050.

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

2026-04-10 21:38:14 +02:00

912 B

Raw Blame History

View Raw